openstackgerrit | Anil Belur proposed openstack-infra/jenkins-job-builder master: Add support for plugin column views https://review.openstack.org/538848 | 02:55 |
---|---|---|
openstackgerrit | Anil Belur proposed openstack-infra/jenkins-job-builder master: Add support for plugin column views https://review.openstack.org/538848 | 03:14 |
openstackgerrit | Anil Belur proposed openstack-infra/jenkins-job-builder master: Add last-stable to default view columns https://review.openstack.org/539820 | 05:50 |
openstackgerrit | Tomas Janousek proposed openstack-infra/jenkins-job-builder master: tests/base: Use plugins_info in SingleJobTestCase as well https://review.openstack.org/539510 | 10:44 |
openstackgerrit | Tomas Janousek proposed openstack-infra/jenkins-job-builder master: hipchat_notif: Fix compatibility dispatch https://review.openstack.org/539511 | 10:44 |
openstackgerrit | Tomas Janousek proposed openstack-infra/jenkins-job-builder master: tests/base: Drop "is not None" https://review.openstack.org/539898 | 10:44 |
*** zaro_ has joined #openstack-jjb | 11:08 | |
*** zaro has quit IRC | 11:11 | |
ssbarnea | does anyone knows if is possible to create a job in a folder without changing its name from foo to folder/foo? (like using a param to specify the folder instead?) | 11:45 |
ssbarnea | i am asking this because this doesn't play well with templates | 11:45 |
*** rubbish has joined #openstack-jjb | 11:58 | |
*** rubbish has quit IRC | 12:16 | |
Odd_Bloke | I haven't used folders, I'm afraid. | 12:44 |
*** rubbish has joined #openstack-jjb | 14:24 | |
rubbish | Hello, I am trying to execute the jjb tests. When running 'tox -e py34' I get the error "'OS_STDOUT_CAPTURE' is not recognized as an internal or external command'..." | 14:25 |
rubbish | It ends with "ERROR: InvocationError: '<mypath>\\jenkins-job-builder\\.tox\\py34\\Scripts\\python.EXE setup.py testr --slowest --testr-args='" | 14:25 |
rubbish | I cannot find any clue on what might be wrong with the setup. | 14:26 |
rubbish | Using Windows 10 with python 3.4.4 | 14:26 |
Odd_Bloke | rubbish: It's possible that they've never worked on Windows; I don't know for sure that any of the developers currently use Windows. | 14:26 |
rubbish | ok, that helps | 14:27 |
rubbish | Trying a VM then, thanks. | 14:27 |
Odd_Bloke | rubbish: If you could paste a more full traceback/log, that would be helpful. | 14:29 |
rubbish | 0 rstein@rstein-PC /cygdrive/r/developer/puppet/jenkins-job-builder $ tox -e py34 py34 develop-inst-nodeps: R:\developer\puppet\jenkins-job-builder WARNING:Discarding $PYTHONPATH from environment, to override specify PYTHONPATH in 'passenv' in your configuration. py34 installed: alabaster==0.7.10,Babel==2.5.3,certifi==2018.1.18,chardet==3.0.4,colorama==0.3.9,coverage==4.4.2,docutils==0.14,extras==1.0.0,fasteners==0.14.1,fixtu | 14:29 |
rubbish | hrm, sorry | 14:30 |
Odd_Bloke | rubbish: Using a pastebin would be preferable. :) | 14:30 |
rubbish | coming ... | 14:30 |
Odd_Bloke | :) | 14:30 |
rubbish | https://pastebin.com/g4mhrAHQ | 14:32 |
ssbarnea | another interesting question is what is the status of jinja2 support? | 14:37 |
ssbarnea | from what i see is not possible to use jinja2 templating for parsing a groovy script to the dsl parameter. | 14:43 |
ssbarnea | somehow it seems to work only with yaml, files.... if I am not missing something. | 14:43 |
Odd_Bloke | ssbarnea: You have to use !include-jinja2: for Jinja2 to be used. | 14:53 |
ssbarnea | Odd_Bloke: this is what I tried and failed to succeed, even with an empty jinja2 file. | 14:54 |
Odd_Bloke | ssbarnea: Could you paste a basic reproducer somewhere? | 14:55 |
ssbarnea | Odd_Bloke: yep, this is what i am trying not to produce. i suspect maybe my checkout is broken or outdated. | 14:55 |
rubbish | Odd_Bloke: Thx - Can work now using stupid Vagrant file. ;) | 15:00 |
Odd_Bloke | rubbish: :) | 15:00 |
ssbarnea | Odd_Bloke: here is basic example https://github.com/ssbarnea/jjb-test | 15:01 |
ssbarnea | i put the error in issue #1, but I guess it would not be hard to reproduce, i did run it on jjb master. | 15:03 |
Odd_Bloke | Huh, that's weird. | 15:06 |
Odd_Bloke | ssbarnea: I'm in a bunch of meetings today, but I'll try to find time to dig in. | 15:21 |
ssbarnea | Odd_Bloke: thanks, no pressure right now, this was the first attempt to use jinja2 in jjb. just ping me when you have something. | 15:22 |
Odd_Bloke | I can reproduce that traceback locally, at least. :) | 15:25 |
Odd_Bloke | OK, that example also fails on the commit that introduced Jinja2 stuff, which we've been using in production for ages. | 15:31 |
Odd_Bloke | So that's very weird. | 15:31 |
ssbarnea | Odd_Bloke: probably because nobody is using jinja2 (yet)? | 15:32 |
Odd_Bloke | Nope, we're using it extensively. | 15:32 |
ssbarnea | ahh, that's good for me, big incentive to have it fixed :D | 15:32 |
Odd_Bloke | Well, it's also not broken for us. :p | 15:33 |
rubbish | Regarding the test suite of jjb - are there any currently failing tests or should everything be green? | 16:07 |
Odd_Bloke | ssbarnea: Aha, I think we only use it within job-templates, modifying your example to use one fixes it. | 16:09 |
Odd_Bloke | So that at least narrows down the scope of the problem. | 16:09 |
Odd_Bloke | (And explains why we aren't seeing this issue elsewhere. :p) | 16:09 |
ssbarnea | thanks | 16:10 |
Odd_Bloke | ssbarnea: And, actually, this is a more general problem; changing !include-jinja2 to !include also produces a very similar traceback. | 16:13 |
Odd_Bloke | I'm bisecting now. | 16:21 |
Odd_Bloke | HAH, the commit to blame is mine. | 16:23 |
Odd_Bloke | But only because I switched the shell builder to use convert_mapping_to_xml. | 16:24 |
Odd_Bloke | OK, so I'm not sure how to handle this. | 16:35 |
Odd_Bloke | OK, the traceback only happens with !include if the file is empty. | 16:37 |
Odd_Bloke | But it continues to happen with !include-jinja2 if it isn't. | 16:37 |
Odd_Bloke | Still not sure how to handle this. :p | 17:01 |
Odd_Bloke | The problem is basically that we don't do the template expansion on non-job-templates (obviously). | 17:01 |
Odd_Bloke | Which means that we don't ever format the Jinja2Loader in to a string. | 17:01 |
Odd_Bloke | Which means that it's assumed to be a dictionary by convert_mapping_to_xml. | 17:02 |
*** openstackgerrit has quit IRC | 18:18 | |
*** caphrim007 has joined #openstack-jjb | 19:28 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!